1 |
haven't seen this on the list as yet and so i'm reposting .. |
2 |
|
3 |
----- Forwarded message from Calum Selkirk <cselkirk@g.o> ----- |
4 |
|
5 |
> Date: Tue, 8 Oct 2002 11:53:35 +0200 |
6 |
> From: Calum Selkirk <cselkirk@g.o> |
7 |
> Reply-To: Calum Selkirk <cselkirk@g.o> |
8 |
> To: gentooppc-user@g.o |
9 |
> Subject: Re: Gentoo-ppc Newbie Questions |
10 |
|
11 |
> * Marcio Luis Teixeira [marciot@×××××××××××××××.edu] [2002-10-06 16:16 -0600]: |
12 |
> |
13 |
> > I've recently successfully got Gentoo PPC running on my old PowerBook |
14 |
> > 3400c. It compiled for five days straight (whew!) but eventually I |
15 |
> > had a working system. It's looking good so far and seems to work |
16 |
> > better than other precompiled distributions (my system now goes to |
17 |
> > sleep without crashing [..] |
18 |
> |
19 |
> this would be the pll, mclk, issue (with older kernels not establishing the |
20 |
> correct values for the ATI card). This hit other Powerbook models also |
21 |
> (though much less than it hit the 3500). Sometime in the not do distant |
22 |
> past atyfb was revamped and so there is no longer the need for users to |
23 |
> pass pll and mclk values to the kernel (via kernel args). The 3500 not |
24 |
> crashing on sleep is the upshot of this. |
25 |
> |
26 |
> > I apologize for posting all this to the mailing list. I'm sure there |
27 |
> > are more appropriate places for each of these items, but I don't yet |
28 |
> > know what they are. If anything, I would appreciate it if you could |
29 |
> > point me in the right direction :) |
30 |
> |
31 |
> heh .. no no .. here is definitely an appropriate place (in fact, I for |
32 |
> one would like to see more traffic on this list in general, in order to |
33 |
> build up some kind of knowledge base). |
34 |
> |
35 |
> [..] |
36 |
> > 1) The docs for the PPC Install say that to enable mouse button |
37 |
> > emulation you must enter into /etc/conf.d/local.start the following |
38 |
> > lines: |
39 |
> > |
40 |
> > echo "1" > /proc/sys/dev/mac_hid/mouse_button_emulation |
41 |
> > echo "87" > /proc/sys/dev/mac_hid/mouse_button2_keycode |
42 |
> > echo "88" > /proc/sys/dev/mac_hid/mouse_button2_keycode |
43 |
> > |
44 |
> > That wasn't working for me, so I changed the third line to |
45 |
> > "mouse_button3_keycode". Seems to work now, so I'm thinking that this |
46 |
> > may be a bug in the docs. Is it? |
47 |
> |
48 |
> yes .. but in more ways than you might think. The whole concept of |
49 |
> echoing digits into proc is depreciated in favor of sysctl (see: |
50 |
> /usr/src/linux/Documentation/sysctl/README). All these values should |
51 |
> be set up via a /etc/sysctl.conf (and note we *do* CONFIG_SYSCTL=y |
52 |
> for this very reason). |
53 |
> |
54 |
> This was reported as a bug to rajiv some weeks .. see: |
55 |
> |
56 |
> http://bugs.gentoo.org/show_bug.cgi?id=8118 |
57 |
> |
58 |
> An /etc/sysctl.conf with the following values will achive the same |
59 |
> effect. |
60 |
> |
61 |
> # Enable mouse button emulation |
62 |
> dev.mac_hid.mouse_button_emulation = 1 |
63 |
> # Set 2nd button to 87 - F11 |
64 |
> dev.mac_hid.mouse_button2_keycode = 87 |
65 |
> # Set 3rd button to 88 - F12 |
66 |
> dev.mac_hid.mouse_button3_keycode = 88 |
67 |
> |
68 |
> > 2) In the past, to get PCMCIA to work on my PowerBook, I had to change |
69 |
> > the default settings in /etc/pcmcia/config.opts as such: |
70 |
> > |
71 |
> > # Settings for the PowerBook: |
72 |
> > include port 0x100-0x4ff, port 0x1000-0x17ff |
73 |
> > include memory 0x80000000-0x80ffffff |
74 |
> > |
75 |
> > This is directly from the PCMCIA How-To: |
76 |
> > |
77 |
> > http://pcmcia-cs.sourceforge.net/ftp/doc/PCMCIA-HOWTO-2.html#ss2.5 |
78 |
> > |
79 |
> > I'm wondering if a) this still applies to Gentoo, and if so b) whether |
80 |
> > this should be part of the installation guide (at least for PowerBook |
81 |
> > users). |
82 |
> |
83 |
> These are the default values when you install pcmcia-cs on gentoo. |
84 |
> |
85 |
> zsh% cat /etc/pcmcia/config.opts | egrep "^inc" |
86 |
> include port 0x1000-0x1fff |
87 |
> include memory 0x90000000-0x9003ffff |
88 |
> |
89 |
> The presence of two ports in the PCMCIA-HOWTO suggest a 3500 or Wallstreet |
90 |
> (whereas the defaults seem to be targeting more recent models). If these |
91 |
> worked for you in the past i would suggest sticking with them (and |
92 |
> please report how you get on). |
93 |
> |
94 |
> > 3) I had a hell of a time figuring out how to configure my mouse under |
95 |
> > XFree86. Would it make sense to add a note about this to the PPC |
96 |
> > install instructions? Basically, I had to set the mouse protocol to |
97 |
> > "PS/2" and the device to "/dev/input/mice." This is definitely not |
98 |
> > what I had expected, as in other distros I had used "BusMouse" and |
99 |
> > "/dev/mouse". Again, I wonder if the PPC Install guide should include |
100 |
> > such info for PowerBook users. |
101 |
> |
102 |
> Sometime around 2.2.17 there were changes make to the (ppc) kernel |
103 |
> known infamously as "the new input layer" or "mac_hid". These changes |
104 |
> depreciated the old ADB keycodes in favor of linux keycodes. However |
105 |
> the old method was (and still is) supported. The upshot of this was lots |
106 |
> of chickens running around all using different keycodes and clucking |
107 |
> furiously on the various list of the time about what needed to be |
108 |
> enabled/disabled in the kernel, what to echo in to /proc and most |
109 |
> famously "my keyboard is b0rked". Gentoo took a rather novel approach to |
110 |
> this (though we still encountered some users who, using non-gentoo |
111 |
> kernels, also exclaimed "b0rked") by disabling CONFIG_MAC_ADBKEYCODES |
112 |
> altogether, and gerk wrote Xeasyconf which would set the option "Protocol" |
113 |
> to "IMPS/2" in /etc/X11/XF86Config-4. Of course the situation with the |
114 |
> new input layer is still not altogether behind us, but we have done our |
115 |
> best to ensure that users don't encounter this (using XF86Config's from |
116 |
> other installs aside). Something for the doc's? .. maybe .. but we can |
117 |
> not cover every piece of legacy code out there, and IMO it's best we |
118 |
> move on. |
119 |
> |
120 |
> > 4) Once I do "kdm" to start the window manager, I am able to log in |
121 |
> > correctly. But, once I log out, the mouse stops working and freezes. |
122 |
> > I have to hit "Alt-E" to restart the X server before logging in again. |
123 |
> |
124 |
> ALT-E? (maybe you mean ALT-c?) .. no idea about this, i avoid {k,g,x)dm |
125 |
> like the plague ;) |
126 |
> |
127 |
> Tip from an old shell geezer: |
128 |
> |
129 |
> In your .bash_profile or .zprofile (dependent on what your shell is). |
130 |
> |
131 |
> [ -z "$DISPLAY" ] && [ -n `tty | grep tty` ] && exec startx |
132 |
> |
133 |
> Then you can skip kdm, and fancy splash screens altogether (unless |
134 |
> there is some initialization kdm performs that i'm unaware of). |
135 |
> |
136 |
> > 5) My CD-ROM drive does not work. At startup I get: |
137 |
> > |
138 |
> > hdc: packet command error: status=0x51 { DriveReady SeekComplete Error } |
139 |
> > hdc: packet command error: error=0x54 |
140 |
> > ATAPI device hdc: |
141 |
> > Error: Illegal request -- (Sense key=0x05) |
142 |
> > Invalid field in command packet -- (asc=0x24, ascq=0x00) |
143 |
> > The failed "Start/Stop Unit" packet command was: |
144 |
> > "1b 00 00 00 03 00 00 00 00 00 00 00 " |
145 |
> > cdrom: open failed. |
146 |
> |
147 |
> humm .. doesn't work at all? How did you install? Are you able to mount |
148 |
> CD's? What errors do you get when mounting? Does /dev/cdroms/cdrom0 |
149 |
> exist? |
150 |
> |
151 |
> > In other distos, I think I had to use SCSI emulation for this drive, |
152 |
> > so I think that may be my problem (still haven't had the time to |
153 |
> > research how to do this under Gentoo). |
154 |
> |
155 |
> |
156 |
> > Grief with Emerge: |
157 |
> > -------------------- |
158 |
> > |
159 |
> > 6) I was able to successfully emerge "alsa-driver" and "alsa-lib", but |
160 |
> > when I try emerging "alsa-utils" (so I can use "amixer" to set the |
161 |
> > volume, as per the desktop configuration guide), I get: |
162 |
> > |
163 |
> > emerge: all ebuilds that could satisfy "alsa-utils" have been masked. |
164 |
> > |
165 |
> > What gives? |
166 |
> |
167 |
> The package is masked for testing (see: /usr/portage/profiles/package.mask) |
168 |
> |
169 |
> It is possible to override this by specifying the full path of the ebuild: |
170 |
> |
171 |
> emerge /usr/portage/media-sound/alsa-utils/alsa-utils-0.9.0_rc3.ebuild |
172 |
> |
173 |
> > 7) When I try emerging "sketch", the emerge fails. Here's what I get: |
174 |
> > |
175 |
> > >>> emerge media-gfx/sketch-0.6.12-r1 to / |
176 |
> > ... blah, blah ... |
177 |
> > >>> md5 ;-) sketch-0.6.12.tar.gz |
178 |
> > >>> Unpacking source... |
179 |
> > >>> Unpacking sketch-0.6.12.tar.gz |
180 |
> > >>> Source unpacked. |
181 |
> > looking for include dir for Imaging.h under /usr/include/python2.2 |
182 |
> > Imaging.h not found under /usr/include/python2.2 ! |
183 |
> > |
184 |
> > !!! ERROR: The ebuild did not complete successfully. |
185 |
> > !!! Function src_compile, Line 27, Exitcode 1 |
186 |
> > !!! setup.py configure failed |
187 |
> > |
188 |
> > I guessed that perhaps "sketch" depended on the python Imaging |
189 |
> > library, so I tried "emerge dev-python/Imaging" and I get the dreaded |
190 |
> > masked message: |
191 |
> > |
192 |
> > merge: all ebuilds that could satisfy "dev-python/Imaging" have been |
193 |
> > masked. |
194 |
> |
195 |
> specify the full path to dev-python/Imaging-1.1.2-r2.ebuild and try |
196 |
> again to merge sketch. Imaging-1.1.2-r2.ebuild is not KEYWORDed ppc. |
197 |
> |
198 |
> best |
199 |
> |
200 |
> cal |
201 |
|
202 |
----- End forwarded message ----- |